IAM1203- CEO Extracts Actionable Insights from Massive Datasets

Keshav Pingali is the CEO and co-founder of Katana Graph, a high-performance scale-out graph processing, AI, and analytics company, that extracts actionable insights from massive unstructured datasets. Keshav holds the W.A.”Tex” Moncrief Chair of Computing at the University of Texas at Austin and is a Fellow of the ACM, IEEE, and AAAS.

• CEO Hack: (i) Ability to say no (ii) Doing my duty
• CEO Nugget: Running a high-tech company is more like a chain of responsibilities
• CEO Defined: Duty and responsibility

Website: https://katanagraph.com/

Twitter: https://twitter.com/KatanaGraph
LinkedIn: https://www.linkedin.com/company/katana-graph/

Check out one of our favorite CEO Hack’s Audible. Get your free audiobook and check out more of our favorite CEO Hacks HERE

Transcription

The full transcription is only available to CBNation Library Members. Sign up today!

Please Note: Our team is using the AI CEO Hacks: Exemplary AI and Otter.ai to support our podcast transcription. While we know it's improving there may be some inaccuracies, we are updating and improving them. Please contact us if you notice any issues, you can also test out Exemplary AI here.

00:52 – Intro

Are you ready to hear business stories and learn effective ways to build relationships, generate sales, and level up your business from awesome CEOs, entrepreneurs, and founders without listening to a long, long, long interview? If so, you've come to the right place. Gresh values your time and is ready to share with you the valuable info you're in search of. This is the I AM CEO podcast.

01:19 – Gresham Harkless

Hello. Hello. Hello. This is Gresh from the I AM CEO podcast. I have a very special guest on the show today. I have Keshav Pingali of Katana graphite. Keshav, it's great to have you on the show.

01:28 – Keshav Pingali

It's good to see you. Gresh, thank you so much for having me on your show.

01:32 – Gresham Harkless

Yes, definitely. The pleasure is all ours. You're doing so many phenomenal things, and what I wanted to do is just read a little bit more about Kishav so you can hear about all the awesome things that he's doing. Kishav is the CEO and co-founder of Katana Graph, a high-performance, scale-out graph processing AI analytics company, that extracts actionable insights from massive unobstructed data sets. Kashaf holds the Waddenne Tex Moncrief chair of computing at the University of Texas at Austin and is a fellow of the ACM, IEE, and AAAS. Keshav, great to have you on the show I AM CEO over here. Are you ready to speak to the I AM CEO community?

[restrict paid=”true”]

02:10 – Keshav Pingali

Yes, I am.

02:11 – Gresham Harkless

Awesome. Yes. Well, thank you for doing all the hard work so I get the opportunity to introduce you. So what I wanted to do to kind of kick everything off was rewind the clock a little bit, hear a little bit more on how you got started, what I call your CEO story.

02:24 – Keshav Pingali

Sure. So I grew up in India. I went to MIT for my PhD, and then I was a professor at Cornell for several years. Then I moved to UT Austin, where I'm a chair, and professor so my entire career, I've been working in high-performance computing and traditionally high-performance computing, parallel computing, cluster computing, has been associated with computational science applications. So simulations of aircraft wings and things like that. But about ten years ago, I and a group of my colleagues in high-performance computing realized that there was a big opportunity to bring techniques that we had developed for computational science applications and bring them to bear on graph analytics, and graph AI, obtaining insights very quickly from graph data.

So I've been doing that for the past ten years. We did a bunch of DARPA projects because DARPA was interested in this technology, and we built a system for them that did real-time intrusion detection in computer networks using the graph technology, the graph platform that we built. DARPA liked it sufficiently that they encouraged us to do a startup. About that time, when this DARPA project ended, about two years ago, I ran into a couple of very well-known entrepreneurs in Silicon Valley. So Libertan and Amadji Gill, and asked me what we were doing. They saw the technology, and they said, well, you got to do a startup. And they said, I'm a professor, I'm doing research. No, no, no, this is the time to do a startup in graph plus AI. And so that's the short version of how we started. Katana, nice.

04:12 – Gresham Harkless

I absolutely love that. It kind of sounds like it gives, I guess, for lack of a better term, people. Those eyes and the ability to kind of see, like you said, intrusions before they actually happen. So I feel like it allows you to kind of like, quote-unquote, see your blind spots, it sounds like as an organization. Is that correct?

04:28 – Keshav Pingali

Yeah. So that's one application of graph knowledge. But what we're building is what we call a graph intelligence platform. So there is, you know, you've heard terms like the data tsunami or the data flood, so the world is generating data at a much faster pace. Then we can analyze the data to gain insights quickly, taking advantage of the window of opportunity. If you can get your insights within that window of opportunity, then you actually benefit from it. But otherwise, if it's too late, well, then it doesn't matter because you can't exploit the insights. So there's the volume of data, and then there is the need to process all of this data very quickly.

So that's where we come in because there is what is called structured data. So structured data is data that you can put into tables, and you have relational databases, that have been around for 50 years. Well-known technology, lots of great companies like Oracle and so on in that area. But increasingly, the data that people are generating, our devices are generating, like, IoT devices. They are what are called linked data. So it's unstructured data that can be represented like a graph and then processed using graph algorithms.

For your listeners who may not be familiar with graphs: everyone uses graphs, even if they are not aware of the term. So if you take an airline route map, for example, what you see are cities, and then lines that correspond to flights between cities. That's a graph. So we call the cities nodes, and then those lines connecting nodes, we call them edges of the graph. Now, an airline route map typically might have 100 or 200 nodes and 200 cities. The graph data we deal with may have billions, hundreds of billions of nodes, and trillions of edges. So there is a big need for processing that kind of enormous scale. Make data and do it very fast.

06:28 – Gresham Harkless

Yeah, that makes so much sense. I appreciate you breaking that down and then especially understanding that time is of the essence in a lot of these situations where you need to be able to understand and know what's going on with that data and be able to kind of, it sounds like, make decisions based off of that.

06:44 – Keshav Pingali

Yeah, that's right. That's right. So it's what we call cognitive intelligence. So, at the simplest level, you take data that comes in and you represent it like a graph, and then you can just say what has happened. Right. At the next level up, you have what is called prescriptive analytics. So, in prescriptive analytics, you also say why something happened. Then you want to get to predictive analytics. That's even more difficult. So there you're trying to predict what may happen in the future based on everything that has happened in the past. Then where we want to be is what we call cognitive analytics.

So there you basically are trying to reason or infer automatically, using AI and machine learning techniques, about why things are happening the way they are. So you build models, as we call them, and then you use the models in order to make predictions about what may happen in the future. So that's where we are, cognitive.

08:15 – Gresham Harkless

Nice. I absolutely love that. Then hearing the application of all this data that we have available to you and how you're able to make that predictive analysis and understand exactly what's potentially going to happen and make those decisions is absolutely huge. I love that you've been able to distill those things. I feel like I'll get an A now in my class, professor, so I appreciate you for breaking all of that down for me. So I know you already touched on a little bit about how serving your clients and how the katana graph works. Is there anything you wanted to add there? Then what you feel is kind of like your secret sauce could be for you individually, or the organization, or a combination of both, and set you apart and make you unique.

08:51 – Keshav Pingali

Yeah, I think what I'd like to say about Katana, what makes Katana unique is if you look at other graph computing companies in this space, they are what we call databases first. So computing is secondary to what they do, which is searching the graph data for certain patterns that you're looking for. So there are companies in this space. What we are focusing on, however, is the analytics and cognitive analytics and the graph intelligence aspect of it, which is not really search, but it focuses much more on very large-scale computing in order to extract properties of these graphs.

So I'll give you a couple of examples from some of our users. Right, the companies we work with. So we have gotten a lot of traction in the medical and pharma area, for example. So we are tied up with one of the biggest pharma companies in the world.

So they have very big graphs that are called knowledge graphs. So these knowledge graphs basically incorporate everything that's known in the medical world, and it's represented as a big graph over there. So the vertices or the nodes of this graph might represent people, papers, biological organisms, and treatments, and then the edges correspond to relationships between all of these entities. Their aim is to mine these knowledge graphs to conduct what they refer to as hypothesis generation.

So hypothesis generation basically finds promising treatments for some particular disease, and they also want to be able to rule out certain potential treatments, because it's a lot faster and cheaper to do it using a computer than going to a wet lab and then finding out that something doesn't work.

So that's an example that we're actively working in, and there are lots of medical companies that we could work with. Another area that is very exciting is what is called precision medicine. So if you go to a doctor right now and he diagnoses you with some disease, chances are he'll give you some treatment that's very similar to what he gives everybody else who has that disease.

But obviously, if we take into account your medical history and your genetic profiles and everything that's known about you, parents and grandparents, all the information that makes you you, makes you special, then we might be able to give you a more targeted treatment than the generic treatment.

So, again, there is an enormous amount of data that needs to be processed fast. So we're doing another contract with a couple of precision medicine companies. Then there are other applications in security. We also already talked about intrusion detection in computer networks.

Bad guys are trying to break into a network and, you know, misuse your resources. You want to catch them as quickly as possible. So graph technologies are used very heavily over there. Again, the graphs are very big. In that application, you can see why time is very important because you want to catch the bad guys before they have a chance to do any damage, lasting damage to your information security.

So there are lots and lots of applications of graph technologies, and in particular, this graph intelligence platform that we are building. Again, just to sort of tie this back, what makes us unique is the fact that we are what we call a graph intelligence platform first, as opposed to a graph database first. So we focus on the compute rather than on the search capability, which is where traditionally graph companies are focused on.

12:22 – Gresham Harkless

Would you consider that to be what I like to call your CEO hack? The thing that makes you more effective and efficient is the ability to say no to sometimes all the opportunities that you might be presented with, but decide to laser focus on those three verticals that you have.

12:34 – Keshav Pingali

Yeah, I think that's one of them. So the ability to say no, or the discipline, perhaps, to say focus on others, I think I found that to be very important. It was one piece of advice that our investors I was telling you about, Libutan and Amerjeet Gill, gave me when we got started, they said that is the number one thing that you have to remember. Then the other thing that helps me in terms of a business hack or a business resource is, you know, I grew up, in India, as I was telling you, and one of our ancient books is called the Bhagavad Gita.

So the Bhagavad Gita is, you can think of it as sort of like one of the books of the Bible, right, in a Western context. What the Gita tells us is that the most noble thing any human being can do is to do his duty, right? So doing your duty is the most important thing. If you've done your duty well, then you've fulfilled your obligations as a human being.

13:37 – Gresham Harkless

No, I absolutely appreciate that. So now I wanted to ask you my absolute favorite question, which is the definition of what it means to be a CEO. We're hoping to have different, quote-unquote, CEO's on this show. So, Keisha, what does being a CEO mean to you?

13:49 – Keshav Pingali

I think I'd go back to the answer I gave you earlier about duty and responsibility. Right. So there are three constituents for me as a CEO, three groups that I have a duty to, a responsibility to. First and foremost are customers because they trust us to solve problems for them. Equally important is my responsibility to my team. So we try to have a nurturing culture over here. One of our mottos at Katana is this Japanese word called kaizen, which means continuous improvement.

So you shouldn't be afraid of making mistakes, but you shouldn't make the same mistake twice because otherwise, if you're punitive about people who make mistakes, then nobody will take risks. That is the biggest risk of all. So we nurture our team members. We help them to grow in their careers. Then the third constituency, of course, is our investors because they have put a lot of trust and, frankly, money in Katana.

We need to make sure that we are responsive to their imperatives. So I think for me, being a CEO just goes back to these two words, duty and responsibility. Then the final thing that I'd like to add as an immigrant to the US is that I have benefited tremendously from the openness of American society, and the ability of Americans to judge people based on their talents and what they bring to this country as opposed to where they were born.

And, of course, we need to make the country a better place, and we need to eliminate problems that exist and so on. But I am very grateful as an immigrant for the opportunities that America has given me. I want the rest of my career to put back into America and make this a better place for everybody who was born here as well as people who move here from other places.

15:43 – Gresham Harkless

I appreciate you so much for doing that and that perspective. Of course, you know, for your time today, one thing I wanted to do was pass you the mic again, just see if there's anything additional that you want to let our readers and listeners know and, of course, how best they can get ahold of you and find about all the awesome things you entertain.

15:57 – Keshav Pingali

You know, I just want to go back to our main theme, which is we're building this graph intelligence platform, scaling out very fast processing of enormous amounts of data to get actionable insights. So I think that is the wave of the future, and we plan to be at the crest of that wave. You can get in touch with me, can go to our website, which is katanagraph.com, and you'll find a lot more information about Katana, what we do, team members, customers, and all of that.

We also have a presence on Twitter, LinkedIn, and YouTube, and all those links are there on our website. Of course, you can send me an email. So, kpingali@catanograph.com will always reach me.

16:43 – Gresham Harkless

Awesome. Awesome. Awesome. To make it even easier, we'll have the links and information in the show notes too, so that everybody can follow up with you. But he Shabbat, I truly appreciate, you know, that, that, um, you know, that, that spirit that you have and that you brought to us today and to remind us of how important it is to constantly get better, whether we're talking about the product that you're creating or the things that we could do related to our duty, our responsibility and understanding the, the people that we have an impact in so many different ways. So thank you so much, for reminding us of that. Of course, doing that so well, and I hope you have a phenomenal rest of the day.

17:11 – Outro

Thank you for listening to the I AM CEO Podcast powered by Blue 16 Media. Tune in next time and visit us at iamceo.co I AM CEO is not just a phrase, it's a community. Be sure to follow us on social media and subscribe to our podcast on iTunes Google Play and everywhere you listen to podcasts, SUBSCRIBE, and leave us a five-star rating grab CEO gear at www.ceogear.co. This has been the I AM CEO Podcast with Gresham Harkless. Thank you for listening.

[/restrict]